| cpu_info | Inspect CPU resources visible to the current process |
| cuda_build_info | Generate CUDA build architecture information |
| cuda_info | Inspect CUDA hardware and system components |
| cuda_libraries | Inspect NVIDIA CUDA shared libraries |
| cuda_preflight | Diagnose CUDA prerequisites |
| gpu_devices | List detected GPU devices |
| gpu_memory_info | Report total, used, and free GPU memory |
| gpu_memory_peak | Calculate observed peak process memory for each GPU |
| gpu_memory_sample | Sample process-level GPU memory |
| gpu_sitrep | Create a GPU diagnostic report |
| hardware_info | Collect hardware information |
| hardware_manifest | Create a reusable hardware manifest |
| has_cuda | Test whether CUDA appears usable |
| has_gpu | Test whether a GPU is detected |
| has_metal | Test whether Metal appears usable |
| has_opencl | Test whether OpenCL is available |
| has_rocm | Test whether ROCm appears usable |
| nccl_info | Inspect the NCCL shared library |
| system_memory_info | Inspect host and process memory |